A 3D visualization‐based augmented reality application for brain tumor segmentation

• Published in: Computer animation and virtual worlds Vol. 35; no. 1
• Main Authors: Guerroudji, Mohamed Amine, Amara, Kahina, Lichouri, Mohamed, Zenati, Nadia, Masmoudi, Mostefa
• Format: Journal Article
• Language: English
• Published: Chichester Wiley Subscription Services, Inc 01.01.2024
• Subjects: 3D reconstruction; 3D visualization; active contour; Augmented reality; Brain; Brain cancer; brain tumor; Cancer; Heterogeneity; Image reconstruction; interaction; Magnetic resonance imaging; MRI; Preprocessing; segmentation; Tumors; Visualization
• ISSN: 1546-4261; 1546-427X
• DOI: 10.1002/cav.2223

Summary Every year on June 8th, the globe observes World Brain Tumor Day to raise awareness and educate people about brain cancer, encompassing both noncancerous (benign) and cancerous (malignant) growths. Research in the field of brain cancer plays a vital role in supporting medical professionals. In this context, augmented reality (AR) technology has emerged as a valuable tool, enabling surgeons to visualize underlying structures and offering a cost‐effective and time‐efficient alternative. Our study focuses on the efficient segmentation of brain tumor classes using Magnetic Resonance Imaging (MRI) and incorporates a three‐stage approach: preprocessing, segmentation, and 3D reconstruction & AR display. In the preprocessing stage, a Gaussian filter is applied to mitigate intensity heterogeneity. Segmentation and detection are achieved using active geometric contour models, complemented by morphological operations. To establish 3D brain tumor reconstruction, a genuine scene is virtually integrated using 3D Slicer software. The proposed methodology was validated using a genuine patient dataset comprising 496 MRI scans obtained from the local Bab El Oued university hospital center. The results demonstrate the effectiveness of our approach in achieving accurate 3D brain tumor reconstruction, efficient tumor extraction, and augmented reality visualization. The obtained segmentation results showcased an impressive accuracy of 98.61%, outperforming existing state‐of‐the‐art methods and affirming the efficacy of our proposed strategy. AR‐assisted 3D brain tumor neurovisualisation.